|
Statistics and the US elections
 There is an old statistician joke: "If you can't count the horses then count the hooves and divide by four". This joke is about a classical method used by statisticians: when you can't measure something directly you use indicators: values correlated with what you are trying to measure. For instance sales of horsehoes, saddles, bridles, and al if you need to know the number of horses. With carefully selected indicators you can have a good idea of the number you are trying to estimate and a still better idea of its evolution.So the MSM are saying Biden will win in a landslide. Then why is the attendance at his rallies so low? Why are there so few "Biden" signs in front of houses? Why is he been nuked by Trump in both? Attendance at rallyes and signs in front of houses are not votes but they are correlated with. Now I will play devil's statistician: perhaps it is that Republicans tend to live in houses with a lawn you can put a "Trump" sign on it and Democrats live in appartments so direct comparisons can be misleading. So let's compare Biden to Clinton in 2016 and again Biden is blown out of the water. In fact he is doing worse than Mondale in 1984. Now people not enthousiastic about Biden would not bother about putting a Biden sign and, because of TDS, still vote for him in higher numbers than they did for Clinton. That would hold if there were, say 10% less pro-Biden than pro-Clinton signs but from what people living in America tell me there were 10 Clinton signs for 1 Biden sign. I really don't believe Biden can do better than Clinton did. Not with such abysmal differences in attendance at rallyes and signs in front of houses. I really believe he will do far worse Without ballot stuffing of mexican proportions I don't see how Biden could win. Posted by: JFM 2020-10-13 |
|
| http://www.rantburg.com/poparticle.php?ID=584655 |